Before Smith, Clement, and Wilson, Circuit Judges. Jerry E. Smith, Circuit Judge: This case requires us to determine whether a plaintiff can relitigate a takings claim in federal court that was previously dismissed in state court. Tejas Motel, L.L.C. (“Tejas”)—the repeat litigant in question—owns and operates a small motel of the same name in Mesquite, Texas. The City of Mesquite (“City”) enacted a series of strict zoning ordinances that turned the motel property into a nonconforming use. It then ordered Tejas to comply with the minimum zoning requirements or cease operations. Tejas sued in state court, claiming that the City violated both the state and federal constitutions by effecting a taking without just compensation. The state trial court dismissed. The Texas Court of Appeals affirmed, hold-ing that Tejas’s state claim was procedurally barred and its federal claim was not “viable.” Tejas then brought the same federal takings claim in federal court, seeking a different result. The law of res judicata bars Tejas’s second suit. Because the district court rightly dismissed the case on that ground, we affirm. I. A. Tejas purchased the Tejas Motel property in 2006, but the motel has operated continuously since about 1970. The City’s regulation of hotels has become increasingly restrictive with time. The City first authorized Tejas’s property for use as a motel in 1973 in a general zoning ordinance. Then, in 1997, the City divided its hotels and motels into two categories—”General Service” and “Limited Service” lodging facilities.[1] All lodging facilities that were not General Service were classified as Limited Service, and Tejas claims that all Limited Service facilities were deemed nonconforming uses. The Tejas Motel was one such nonconforming use. In 2008, the City made the definition of a “General Service” facility even more restrictive.[2] Tejas failed to meet those criteria either.[3] After local residents allegedly complained about the poor conditions and criminal activity at nonconforming hotels, the City took further action. In 2018, it revised the process by which it could “amortize” nonconforming establishments.[4] It then asked the municipal Board of Adjustment (the “Board”) to consider forcing five particular nonconforming uses (including Tejas Motel) into compliance with the zoning regulations. The City outlined a two-step process for determining whether to move against Tejas: First, the Board would hold a public hearing to decide whether Tejas’s continued operation would have an adverse effect on nearby properties; second, if so, the Board would set a date by which the motel was required to comply with the zoning regulations. The Board dutifully followed instructions. In July 2018, it gave Tejas notice that it would hold a public hearing to decide whether to amortize the motel. Before the hearing, the City provided Tejas with the evidence it intended to use to prove that the motel’s continued operation was indeed having negative effects on its neighbors. The City also allegedly warned the motel that it was not entitled to any amortization period under the new regulations. Facing pressure from the City and the Board, Tejas agreed to cease operations or bring the motel into conformity with the zoning ordinance by May 1, 2019. The Board issued an order approving the amortization plan. Despite its agreement with the City, Tejas claimed that it settled only under “duress.” Three months later, it sued in state court, claiming that it had suffered a taking in violation of the state and federal constitutions. B. Before analyzing Tejas’s lawsuit, we survey the procedure for bringing a federal takings claim, as the operative rules have changed during the course of the litigation. The Takings Clause of the Fifth Amendment prohibits the government from taking private property “for public use, without just compensation.” U.S. Const. amend. V. That provision is incorporated against state and local governments.[5] But for many years, plaintiffs who suffered a taking at the hands of a local government could not immediately sue in federal court. In its pivotal Williamson County[6] decision, the Supreme Court established what became known as the “state-litigation requirement.”[7] A federal takings claim was not considered “ripe” until a plaintiff obtained “a final decision regarding the application of the zoning ordinance . . . to its property” and “utilized the procedures [the state] provides for obtaining just compensation,” including state judicial process. Williamson Cnty., 473 U.S. at 186, 193. Yet litigation of a takings claim in state court under state law was not merely a procedural precursor to a later hearing in a federal forum. Once a state court made a final adjudication of a takings claim, that judgment had full preclusive effect—relitigation of the claim could be barred by ordinary principles of res judicata. See San Remo, 545 U.S. at 338, 346–48. Commentators quickly realized that the combination of those precedents put litigants in a terrible double-bind.[8] If a plaintiff brought a takings claim in federal court first, the suit would be deemed unripe (per Williamson County). But if he brought a takings claim in state court and lost, res judicata would likely bar any subsequent federal litigation (per San Remo). The Court solved that problem in Knick v. Township of Scott, 139 S. Ct. 2162(2019). It bemoaned the “preclusion trap” created by Williamson County and San Remo, which made the promise of a federal forum for takings plain-tiffs “ring[] hollow.” Id. at 2167. Thus, the Court overruled Williamson County and its “exhaustion requirement.” Id. at 2173, 2179. No longer do plaintiffs need to litigate their takings claims in state court first. Instead, “a government violates the Takings Clause when it takes property without compensation,” and “a property owner may bring a Fifth Amendment claim under § 1983 at that time” in federal court. Id. at 2177. C. In 2018, however, Williamson County was still good law. So when the City decided to amortize Tejas’s property that year, Tejas sued in state court, listing claims under the state constitution and the federal Takings Clause. In response to Tejas’s Third Amended Petition, the City filed a plea to the jurisdiction, a Texas procedural device used primarily to challenge a court’s subject matter jurisdiction. See Tex. Dep’t of Parks & Wildlife v. Miranda, 133 S.W.3d 217, 225–26 (Tex. 2004). The City claimed, among other things, that the court could not hear the suit because the City enjoyed sovereign immunity. The City also filed a motion for partial summary judgment. A hearing on both the plea and the summary judgment motion was scheduled for March 29, 2019. Two days before that hearing, Tejas filed a Fourth Amended Petition, adding the individual Board members as defendants (in their official capacities). The same day, Patricia McCollum—who purported to be a guest at the Tejas Motel—filed a petition in intervention, seeking to enter the lawsuit. Without acknowledging those belated pleadings, the trial court, on May 17, 2019, entered a summary order granting the City’s plea to the jurisdiction in its entirety. A year later, the Texas Court of Appeals affirmed and gave a more detailed rationale. First, it held that Tejas had failed to raise its state law claims in a timely manner. State law required that a challenge to the Board’s amortization plan be brought within ten days using designated state procedures, whereas Tejas waited three months to file a petition in state court. Because Texas’s timeliness requirement was jurisdictional, the City’s plea to the jurisdiction was properly granted on the state takings claim.[9] Second, the court held that “Tejas failed to state any viable federal constitutional claim.” It reasoned that Tejas “has no constitutionally protected interest in continuing to use the property in violation of the City’s zoning ordinances when it acquired the property knowing it was in violation of those ordinances.” Because Tejas failed to state a viable claim, its federal claim was properly dismissed. The Supreme Court of Texas denied review, and the mandate issued to the trial court in December 2020. D. After the decision of the Texas Court of Appeals—but while Tejas’s petition for review was still pending in the state supreme court—Tejas sued in federal court. The complaint was substantially similar to Tejas’s state court filing. The City quickly moved to dismiss. The district court granted that motion on the basis of the Rooker-Feldman doctrine, stating that it could not directly review a state court judgment. Tejas responded with a Federal Rule of Civil Procedure 59 motion to alter or amend the judgment, claiming that Rooker-Feldman was inapplicable because, inter alia, the motel sought relief from the underlying action by the City, not from the state court judgment itself. The district court agreed and confessed error. Nevertheless, it found that Tejas’s claims were still barred by res judicata. It explained that the Texas Court of Appeals had dismissed Tejas’s claims for failing to state a “viable federal constitutional claim.” That dismissal operated as a “final judgment on the merits” and precluded Tejas from re-raising its federal takings claim. The court therefore declined to disturb its original judgment and dismissed Tejas’s case with prejudice. Tejas appeals. E. During Tejas’s federal appeal, there has been some activity back in state court.[10] In a December 21, 2022, order, the district court struck McCollum’s petition in intervention, stating: “This is a full and final order resolving all remaining claims of all remaining parties herein. All relief not expressly granted is hereby denied.” Tejas filed a motion for new trial soon thereafter, contending that the order was in error. Tejas then filed a letter with this court, claiming that its new-trial motion keeps the state court action alive and prevents the application of res judicata in federal court. But the City urges that the December 2022 order has no impact on the federal appeal, given that it dispatches only with McCollum’s attempted intervention. In the City’s view, Tejas’s claims have been resolved conclusively since December 2020. II. We review the grant of a motion to dismiss de novo. Thurman v. Med. Transp. Mgmt., Inc., 982 F.3d 953, 955 (5th Cir. 2020). Dismissal based on res judicata is appropriate under Rule 12(b)(6) where a defendant’s affirmative defense appears “on the face of the complaint.”[11] III. The decisive issue on appeal is whether the adverse judgment in state court precludes Tejas from bringing a federal takings claim against the City. For the answer, we look to Texas law.[12] A claim is barred by res judicata (or claim preclusion) if a defendant can establish “(1) aprior final judgment on the merits by a court of competent jurisdiction; (2) identity of parties or those in privity with them; and (3) a second action based on the same claims as were raised or could have been raised in the first action.” Amstadt v. U.S. Brass Corp., 919 S.W.2d 644, 652 (Tex. 1996). The second and third components are not seriously in dispute. Tejas and the City were both parties in the state court action. And Tejas’s federal takings claim is the same as the one it raised in state court. Instead, all the focus is on the first element. The parties dispute whether the state court judgment was “final,” “on the merits,” and “by a court of competent jurisdiction.” Close analysis proves that each of those requirements is satisfied.[13] A. First, the state court decision was final. To be “final” for purposes of claim preclusion, a judgment must be both comprehensive and definite. A judgment gets preclusive effect when it “settles all rights adjudicated be-tween the parties so that the substantive rights vindicated in it are vested rights”[14] but not “when outstanding issues remain in the action.”[15] Additionally, a final judgment must be the “the ‘last word’ of the rendering court.”[16] A judgment is res judicata once the trial court loses plenary power to amend or alter it.[17] When judged by those criteria, the dismissal of Tejas’s constitutional claims against the City was assuredly final. The state trial court dispatched with all of Tejas’s claims by granting the City’s plea to the jurisdiction, and the Texas Court of Appeals affirmed the dismissal. After Tejas unsuccessfully petitioned for review in the Supreme Court of Texas, the mandate issued to the trial court, effectuating the judgment. At that point, all extant issues between the two parties were conclusively resolved, and the state court had no authority to reverse or alter the judgment in favor of the City.[18] Tejas points out that the appeals court did not make the preclusive effect of its dismissal explicit. The Texas Court of Appeals affirmed the “trial court’s order . . . dismissing Tejas’s claims against the City,” but neither the trial court order nor the appeals court specified whether the dis-missal was with prejudice. Nevertheless, a dismissal is with prejudice where a “court lacks subject matter jurisdiction because of the sovereign immunity bar.” Harris Cnty. v. Sykes, 136 S.W.3d 635, 639 (Tex. 2004).[19] And the Texas Court of Appeals indirectly rested its dismissal on the City’s sovereign immunity. Although a Texas municipality is not generally immune from takings claims, a city retains immunity “[i]n the absence of a properly pled takings claim.”[20] Thus, when a municipality raises a plea to the jurisdiction, courts can grant the plea if the plaintiff’s constitutional claim is not viable. See City of Dallas v. Saucedo-Falls, 268 S.W.3d 653, 657–58 (Tex. App.—Dallas 2008, pet. denied). The Texas Court of Appeals cited Saucedo-Falls to conclude that the plea to the jurisdiction was properly granted because Tejas lacked a “viable . . . takings claim” against the City. Said another way, Tejas’s merits claims were dismissed as nonviable attacks on an immune municipality. Although that posture is unusual, it is a dismissal based on sovereign immunity.[21] Therefore, it is also a dismissal with prejudice.[22] “[A] dismissal with prejudice functions as a final determination” and is preclusive. Mossler v. Shields, 818 S.W.2d 752, 754 (Tex. 1991) (per curiam). Still, Tejas insists that a decision is not final until the “rights of all parties” have been adjudged,[23] and granting the plea to the jurisdiction did not resolve the status of Patricia McCollum (who filed a petition in intervention) and the individual Board members (who were supposedly added as defendants in Tejas’s Fourth Amended Petition and McCollum’s petition). But although the rights of “all” parties often need to be resolved before a judgment is “final” for appeal,[24] a “judgment may be final in a res judicata sense as to a part of an action although the litigation continues as to the rest.”[25] There is no doubt that the state-court dismissal of Tejas’s claims against the City resolved all issues between those parties. It is also undisputed that the state courts are without power to revise that judgment. That makes the judgment final for res judicata. We emphasize that, as a factual matter, the other motions in state court appear to be resolved. First, the Texas Court of Appeals characterized Tejas’s Fourth Amended Petition as untimely.[26] Second, the trial court struck McCollum’s petition in intervention in its December 2022 order. And in case there was any doubt about whether there were lingering motions in state court, the December order clarified that it was a “full and final order resolving all remaining claims of all remaining parties herein.” Tejas correctly points out in its supplemental briefing that the state court could technically still revise the December 2022 order. Because Tejas filed a new trial motion in response to the December 2022 order, the state court retains plenary jurisdiction over the case until the new trial motion is resolved (plus an extra 30 days).[27] Yet the debate over the status of the petition in intervention and the amended pleadings is a red herring—whether or not those filings are on life support does not change our holding. The state court has no power to revise or alter the dismissal of Tejas’s claims against the City.[28] All substantive issues between those parties have been resolved. Therefore, for purposes of res judicata, that judgment is final. B. Additionally, the state court decision was “on the merits.” The Texas Court of Appeals dismissed Tejas’s suit against the City for “fail[ing] to state any viable federal constitutional claim.” In so holding, the court assessed the substance of the constitutional claim and analyzed several precedents before concluding that Tejas had no “vested property interest.” It then relied on the test in Penn Central Transportation Co. v. City of New York, 438 U.S. 104, 124 (1978), to hold that Tejas lacked a “reasonable investment-backed expectation in continuing to operate the motel as a nonconforming use.” It concluded that “Tejas’s federal constitutional claims fail.” By its own terms, that was a merits-based dismissal. Tejas offers two rejoinders, but both fall short. First, Tejas repeatedly emphasizes that the trial court’s original order dismissing the case was ambiguous about the basis for its dismissal. That is true, but the decision of the Texas Court of Appeals was explicitly on the merits. The appellate court judgment is the one that gets preclusive effect here.[29] Second, Tejas insists that its claims against the City were dismissed for lack of subject matter jurisdiction, not on the merits of the constitutional claim. After all, the dismissal was pursuant to a “plea to the jurisdiction.” But reliance on that label is somewhat misleading. As discussed above, a plaintiff may use a plea to the jurisdiction to “challenge the validity” of a federal constitutional claim against a municipality, as government bodies are immune from non-viable claims.[30] And when a court grants a plea to the jurisdiction based on sovereign immunity, that adjudication is treated as “on the merits” for purposes of claim preclusion.[31] That is precisely what happened. The Texas Court of Appeals dis-missed Tejas’s claim for failing to state a viable claim against an immune municipality. Such an order is “on the merits” under Texas’s rules of claim preclusion. C. Tejas spends most of its efforts challenging the subject matter juris-diction of the state court. Its reasoning goes like this: A judgment lacks res judicata effect if it was rendered without jurisdiction. Engelman Irrigation Dist. v. Shields Bros., Inc., 514 S.W.3d 746, 750 (Tex. 2017). Ripeness implicates subject matter jurisdiction. Eagle Oil & Gas Co. v. TRO-X, L.P., 619 S.W.3d 699, 706 (Tex. 2021). At the time of its first lawsuit, Williamson County was still good law—federal takings claims were not ripe “until the entry of a final state judgment denying just compensation.” San Remo, 545 U.S. at 337. Because the state litigation of Tejas’s takings claims had not yet concluded, its federal claim was not ripe. Indeed, because Tejas’s state takings claims were time-barred, Tejas insists that it could never litigate its state claims in state court and thus its federal claim could never ripen. On that theory, the state court lacked jurisdiction over the federal claim and Tejas is free to re-raise it in federal court. Yet Tejas’s claim was ripe for adjudication by the time of the state court appeal. Williamson County‘s state-litigation requirement was overruled in 2019, a year before the Texas Court of Appeals issued its decision. There-fore, at the time the relevant state court judgment, there was no state-litigation requirement. Knick, 139 S. Ct. at 2162, 2179. Tejas’s federal takings claim was ripe under governing federal law, and the state court was competent to adjudicate the dispute. Tejas counters that to be preclusive, a claim must be ripe at the time of filing in state court.[32] Because Tejas believes that its federal claim ripened only post-Knick—that is, between the trial court’s original dismissal and the state appeals court’s affirmance—the motel suggests that its federal claim cannot be barred in a later suit. But even while Williamson County was still good law, Texas allowed federal takings claims to “ripen” during the course of a lawsuit.[33] A plaintiff was permitted to bring federal and state takings claims simultaneously in the same litigation. San Remo, 545 U.S. at 346. Once the state court resolved the takings claim under state law, the federal claim became ripe and could be adjudicated.[34] So long as that procedure was followed, a final judgment on the merits of the federal claim was preclusive.[35] If Tejas’s federal claims could have ripened midway through the lawsuit under the “ancien regime,”[36] then a fortiori its claims ripened before the appellate court when the state-litigation requirement was abrogated. The dismissal of Tejas’s federal constitutional claims was therefore a final decision by a court with subject matter jurisdiction. It meets all three requirements for preclusion under Texas law. D. Tejas has one final response: It claims that Knick was a sea change in takings jurisprudence that prevents the operation of res judicata. It is well established in Texas that “a judgment in one suit will not operate as res judicata to a subsequent suit . . . if there has been a change in . . . the decisional law between the first judgment and the second suit.” Marino v. State Farm Fire & Cas. Ins. Co., 787 S.W.2d 948, 949–50 (Tex. 1990) (underlining re-moved). Because Knick eliminated Williamson County‘s exhaustion requirement after Tejas first filed suit, Tejas reasons that the pre-Knick state court judgment cannot be binding. But that contention fails for three reasons. First, Tejas has the time-line wrong. Although Knick was not on the books when the state trial court first dismissed Tejas’s claims, the case had been decided for a full year before the Texas Court of Appeals ruled in favor of the City. See Knick, 139 S. Ct. at 2162. It was governing law on appeal, and the appeals court’s judgment (and subsequent order) is what gets preclusive effect here. Thus, Knick is not really an “intervening” change between the first and second action, given that it was operative law at the time of the first final judgment. Additionally, even if Knick is “intervening” in a loose sense, Knick had no effect on the merits of Tejas’s claim. Knick eliminated the requirement for takings plaintiffs to exhaust state judicial remedies, but it did not change the substantive law about what constitutes a taking under the federal Constitution. See Knick, 139 S. Ct. at 2177. Said another way, the merits of Tejas’s claim would be adjudicated under the same judicial rules with and without Knick. Finally, Knick did not purport to overrule San Remo, which held that state adjudications of federal takings claims have res judicata effect. 545 U.S. at 346–47. One of our sister circuits has already recognized that pre-Knick adjudications of federal takings claims in state court are still preclusive in the wake of that decision. See Ocean Palm GolfClub P’ship v. City of Flagler Beach, 861 F. App’x 368, 371 (11th Cir. 2021) (per curiam). Although that opinion is unpublished, its analysis is sound. San Remo is still good law. Thus, a state court’s judgment on a federal takings claim issued before Knick is still preclusive after Knick (provided that the elements of res judicata are otherwise satisfied). Because the elements of res judicata are met here, we are bound to respect the state court decision. IV. One can sympathize with Tejas’s procedural plight. It was forced into state court by Williamson County, and now cannot avoid the consequences of the adverse judgment it received. But nothing in Knick nullifies long-settled principles of res judicata. State courts are competent to adjudicate federal claims, and their judgments are entitled to full faith and credit in federal court. See 28 U.S.C. § 1738. Because the Texas Court of Appeals issued a final judgment on the merits of Tejas’s constitutional claim, the motel does not get a second bite at the apple. The judgment of dismissal is AFFIRMED.
