BorlandTalk.com Borland discussion newsgroups   Archives   FAQ   Search   Memberlist   Usergroups   Register   Profile   Log in to check your private messages   Log in  Overlay index in SEGDEF is zero          BorlandTalk.com Forum Index -> C++ Builder (Command Line Tools) View previous topic :: View next topic   Author Message Guest Posted: Thu Jan 25, 2007 11:31 pm    Post subject: Overlay index in SEGDEF is zero This is a bug report: Overlay index values in SEGDEF/SEGDEF32 records the compiler is spitting out are set to zero. This violates original Intel OMF spec and TIS specs. Its already emitting a null LNAMES string (usually as the 3rd LNAME in the list). Using the index of a null LNAME string is the correct way to do this rather than using zero. Back to top Ed Mulroy Guest Posted: Thu Jan 25, 2007 11:33 pm    Post subject: Re: Overlay index in SEGDEF is zero Omitting the version of a program in a problem report limits the value of the report. What version of Borland C++ was used to develop the information in your report? .. Ed Quote: cutaway wrote in message news:45b8e6c8 (AT) newsgroups (DOT) borland.com... This is a bug report: Overlay index values in SEGDEF/SEGDEF32 records the compiler is spitting out are set to zero. This violates original Intel OMF spec and TIS specs. Its already emitting a null LNAMES string (usually as the 3rd LNAME in the list). Using the index of a null LNAME string is the correct way to do this rather than using zero. Back to top Guest Posted: Fri Jan 26, 2007 1:40 am    Post subject: Re: Overlay index in SEGDEF is zero All 32 bit versions after BC 4.52. Basically anything with the newer backend. All record checksums are also bogus always being zero. The 16 bit compilers did correct checksums. "Ed Mulroy" <dont_email_me (AT) bitbuc (DOT) ket> wrote in message news:45b8e9be$1 (AT) newsgroups (DOT) borland.com... Quote: Omitting the version of a program in a problem report limits the value of the report. What version of Borland C++ was used to develop the information in your report? . Ed Back to top Ed Mulroy Guest Posted: Fri Jan 26, 2007 5:31 am    Post subject: Re: Overlay index in SEGDEF is zero A trouble report was publicly acknowledged and posted in 1996 about overlays using Borland C++ 5 and above. The linker incorrectly implemented overlays. To correct it use the linker from BC++ 4.5, a product whose CD shipped with BC++ versions 5.*. Please post Borland C++ questions in the newsgroup where it is covered, borland.public.cpp.borlandcpp As you can see from the 'cppbuilder' in its name, this group is for Borland C++ Builder, the current compiler. It is 32 bit only and does not do overlays. .. Ed Quote: cutaway wrote in message news:45b90504 (AT) newsgroups (DOT) borland.com... All 32 bit versions after BC 4.52. Basically anything with the newer backend. All record checksums are also bogus always being zero. The 16 bit compilers did correct checksums. Back to top Guest Posted: Fri Jan 26, 2007 9:10 am    Post subject: Re: Overlay index in SEGDEF is zero Ummm...this has NOTHING to do with using overlays. This is about bogus index values when you're NOT DOING OVERLAYS. Builder is indicating the NON-PRESENSE of overlays incorrectly. BTW, TASM32 is *NOT* afflicted with this disease, so if zero indexes are be your new (non)standard, then you need to fix (break) TASM32 to make it consistent with the BUILDER product. So which is it to be? Is Builder broken or is TASM32 broken? Pick one. You can't have it both ways. "Ed Mulroy" <dont_email_me (AT) bitbuc (DOT) ket> wrote in message news:45b93f5e (AT) newsgroups (DOT) borland.com... Quote: As you can see from the 'cppbuilder' in its name, this group is for Borland C++ Builder, the current compiler. It is 32 bit only and does not do overlays. . Ed Back to top Guest Posted: Fri Jan 26, 2007 9:10 am    Post subject: Re: Overlay index in SEGDEF is zero Well, the tools I'm producing do. Their whole purpose in life is to work with generic OBJ's. If Borland has chosen to knowingly ignore the TIS standard, this is something that should be stated up front, so 3rd parties like me know what we're getting into should we (perhaps foolishly) try to support Borland's products. Personally, I don't give a damn. Its easy for me to fix this bogus index brokenness in my product and morph the output OBJ into TIS standard compliance. In fact, I've already implemented a -Fi switch that does so in the past couple of days. I also fix your bad checksums, optimize the LINNUM's, LNAMES, PUBDEF's, EXTDEF's, and FIXUPS. I used to work for Borland in the Turbo C group back in the late 80's. It saddens me to see an attitude like yours has apparently taken hold. Back then, we'd have just fixed it rather than justifying and denying it. What happend to the likes of Jervis, Kukol and Stafford? All gone I suppose... "Ed Mulroy" <dont_email_me (AT) bitbuc (DOT) ket> wrote in message news:45b987c4$1 (AT) newsgroups (DOT) borland.com... Quote: ...I don't normally dissasemble binaries. . Ed Back to top Ed Mulroy Guest Posted: Fri Jan 26, 2007 9:10 am    Post subject: Re: Overlay index in SEGDEF is zero Quote: Pick one. You can't have it both ways. Luckily I don't have to pick anything. I use both C++ Builder and Tasm32, often in the same program, and they work. I don't normally dissasemble binaries. .. Ed Quote: cutaway wrote in message news:45b974ee$1 (AT) newsgroups (DOT) borland.com... Ummm...this has NOTHING to do with using overlays. This is about bogus index values when you're NOT DOING OVERLAYS. Builder is indicating the NON-PRESENSE of overlays incorrectly. BTW, TASM32 is *NOT* afflicted with this disease, so if zero indexes are be your new (non)standard, then you need to fix (break) TASM32 to make it consistent with the BUILDER product. So which is it to be? Is Builder broken or is TASM32 broken? Pick one. You can't have it both ways. Back to top vavan Guest Posted: Fri Jan 26, 2007 3:54 pm    Post subject: Re: Overlay index in SEGDEF is zero On Fri, 26 Jan 2007 00:47:06 -0500, <cutaway (AT) bellsouth (DOT) net> wrote: Quote: Well, the tools I'm producing do. Their whole purpose in life is to work with generic OBJ's. I wonder where can I see your tools? Quote: If Borland has chosen to knowingly ignore the TIS standard, this is something that should be stated up front, so 3rd parties like me know what we're getting into should we (perhaps foolishly) try to support Borland's products. well, according to Yury Haron (author of ulink linker) there are many incompatibilities in borland's tools so he had to workaround some of them -- Vladimir Ulchenko aka vavan Back to top Guest Posted: Fri Jan 26, 2007 5:59 pm    Post subject: Re: Overlay index in SEGDEF is zero As a general statement, there's a lot of weird stuff in OBJ's Vlad ;-> I've seen Ryan McFarland Fortran compilers emitting multiple THEADR/MODEND sets in a single OBJ (16 bit TLINK's only see the first set, MS's LINK works OK). I've seen Microfocus COBOL injecting crap zero padding at the end. Microfocus also allows their emitted namespace to contain dashes. Many earlier versions of MASM don't set the "big bit" correctly for exactly 64K segments. Some RM Fortran compilers emit bogus (but not zero) checksums. Microsoft's later 16 bit C/C++ compilers are emitting empty CV info in SEGDEF32's (even when you didn't call for any debug info) - undoubtedly to sabotage using the OBJ's for OS/2 ;-> SLR's OPTASM will violate the charset specs for a THEADR string. MS's BASCOM did something weird I don't remember at the moment... This project has been quite an "adventure". Today I've been working on Borland VIRDEF support and it will probably be done tomorrow sometime. The only major record support missing after that is COMDAT's. I'll send you a snapshot after the VIRDEF stuff goes in. I would like to see it run against a bunch of different compilers and assemblers to see what shakes out. In particular I'd like people to beat on the threaded fixup optimizer/unoptimizer. Some TLINK's behaved rather badly when they encountered threaded fixups, so I put in the fixup "unoptimizer" to remove all threaded fixup references so people can link foreign compiler produced OBJ's using TLINK. On the MS produced OBJ's I've tested against, the threaded fixup "optimizer" does a better job of picking what to thread than the Microsoft compilers did every time. Borland tools never generated threaded fixups, so there's a huge OBJ size win runing the optimizer against them when you link Borland's stuff with some linker that can handle the threaded fixups correctly. I also put in a LINNUM/LINN32 record optimizer that can shrink the OBJ's down considerably for 32 bit code. All the Borland 32 bit backends spit out LINN32's even when the code offets never exceeded 64K. It takes a really large blob of code to have more than 64K in a single OBJ, so that situation happens almost every time and Borland is wasting 2 bytes on every line number subrecord. "vavan" <VavanRemove (AT) ThisSantel (DOT) ru> wrote in message news:tmjjr29i5trqku9v3k6qdqck1qlat7eokg (AT) 4ax (DOT) com... Quote: On Fri, 26 Jan 2007 00:47:06 -0500, <cutaway (AT) bellsouth (DOT) net> wrote: Well, the tools I'm producing do. Their whole purpose in life is to work with generic OBJ's. I wonder where can I see your tools? well, according to Yury Haron (author of ulink linker) there are many incompatibilities in borland's tools so he had to workaround some of them -- Vladimir Ulchenko aka vavan Back to top vavan Guest Posted: Fri Jan 26, 2007 6:37 pm    Post subject: Re: Overlay index in SEGDEF is zero On Fri, 26 Jan 2007 06:59:57 -0500, <cutaway (AT) bellsouth (DOT) net> wrote: Quote: all threaded fixup references so people can link foreign compiler produced OBJ's using TLINK. On the MS produced OBJ's I've tested against, the aforementioned ulink allows "mixed" linking. I personally never needed such capability but you could communicate with Yury if you need well-informed opinion -- Vladimir Ulchenko aka vavan Back to top Ed Mulroy Guest Posted: Fri Jan 26, 2007 10:51 pm    Post subject: Re: Overlay index in SEGDEF is zero Quote: ...I used to work for Borland in the Turbo C group back in the late 80's... We met with the TC developers at Borland during the annual meetings. It is suprising that you do are not aware that no TeamB members are Borland employees. If you did development back then we should have met at least once. I of course cannot know who you are because you are on here cloaked, using an alias instead of your name. Quote: ...It saddens me to see an attitude like yours has apparently taken hold ... You read this wrong. I come here on my own time, to help fellow users. When faced with someone who comes here in assertive mode, I can choose to play or not play his game. I choose to not play. .. Ed Quote: cutaway wrote in message news:45b9932b$1 (AT) newsgroups (DOT) borland.com... Well, the tools I'm producing do. Their whole purpose in life is to work with generic OBJ's. If Borland has chosen to knowingly ignore the TIS standard, this is something that should be stated up front, so 3rd parties like me know what we're getting into should we (perhaps foolishly) try to support Borland's products. Personally, I don't give a damn. Its easy for me to fix this bogus index brokenness in my product and morph the output OBJ into TIS standard compliance. In fact, I've already implemented a -Fi switch that does so in the past couple of days. I also fix your bad checksums, optimize the LINNUM's, LNAMES, PUBDEF's, EXTDEF's, and FIXUPS. I used to work for Borland in the Turbo C group back in the late 80's. It saddens me to see an attitude like yours has apparently taken hold. Back then, we'd have just fixed it rather than justifying and denying it. What happend to the likes of Jervis, Kukol and Stafford? All gone I suppose... Back to top David Dean [CodeGear] Guest Posted: Sat Jan 27, 2007 4:00 am    Post subject: Re: Overlay index in SEGDEF is zero In article <45b8e6c8 (AT) newsgroups (DOT) borland.com>, <cutaway (AT) bellsouth (DOT) net> wrote: Quote: This is a bug report: Overlay index values in SEGDEF/SEGDEF32 records the compiler is spitting out are set to zero. This violates original Intel OMF spec and TIS specs. Its already emitting a null LNAMES string (usually as the 3rd LNAME in the list). Using the index of a null LNAME string is the correct way to do this rather than using zero. http://qc.codegear.com is our bug tracking system. I welcome a good test case. -- -David Dean CodeGear C++ QA Engineer <http://blogs.codegear.com/ddean/> Back to top Guest Posted: Sat Jan 27, 2007 3:25 pm    Post subject: Re: Overlay index in SEGDEF is zero David, I just sent you a nifty toy and some instructions to play with that will reproduce this. "David Dean [CodeGear]" <david.dean.no (AT) spam (DOT) codegear.com> wrote in message news:david.dean.no-B1805F.14001426012007 (AT) killface (DOT) local... Quote: In article <45b8e6c8 (AT) newsgroups (DOT) borland.com>, <cutaway (AT) bellsouth (DOT) net wrote: This is a bug report: Overlay index values in SEGDEF/SEGDEF32 records the compiler is spitting out are set to zero. This violates original Intel OMF spec and TIS specs. Its already emitting a null LNAMES string (usually as the 3rd LNAME in the list). Using the index of a null LNAME string is the correct way to do this rather than using zero. http://qc.codegear.com is our bug tracking system. I welcome a good test case. -- -David Dean CodeGear C++ QA Engineer http://blogs.codegear.com/ddean/ Back to top Display posts from previous: All Posts1 Day7 Days2 Weeks1 Month3 Months6 Months1 Year Oldest FirstNewest First         BorlandTalk.com Forum Index -> C++ Builder (Command Line Tools) All times are GMT Page 1 of 1   Jump to: Select a forum C++ Builder Newsgroups----------------C++ BuilderC++ Builder (ActiveX)C++ Builder (Announcements)C++ Builder (Command Line Tools)C++ Builder DatabasesC++ Builder Databases (ADO)C++ Builder Databases (Desktop)C++ Builder Databases (InterBase Express)C++ Builder Databases (Multi-Tier)C++ Builder Databases (SQL Servers)C++ Builder (Graphics)C++ Builder (IDE)C++ Builder (Internet)C++ Builder (Internet Socket)C++ Builder (Internet Web)C++ Builder (Jobs)C++ Builder (Language)C++ Builder (Language C++)C++ Builder (Compatibility with MFC)C++ Builder (Native API)C++ Builder (Non-Technical)C++ Builder (Students)C++ Builder (Third Party Tools)C++ Builder (Upgrade)C++ Builder (VCL)C++ Builder (VCL Components Usage)C++ Builder (VCL Components Development)C++ Builder (VCL Usage)C++ Builder (WinAPI)C++ Builder Installation Delphi Newsgroups----------------Delphi (General)Delphi ActiveX Controls UsageDelphi ActiveX Controls DevelopmentDelphi As400Delphi DatabasesDelphi Databases (ADO)Delphi Databases (dbExpress)Delphi Databases (Desktop)Delphi Databases (InterBase Express)Delphi Databases (Multi-Tier)Delphi Databases (SQL Servers)Delphi DeploymentDelphi GraphicsDelphi IDEDelphi InternationalizationDelphi InternetDelphi Internet Isapi-WebBrokerDelphi Internet WebSnapDelphi Internet WinsockDelphi JobsDelphi Language BASMDelphi Language ObjectPascalDelphi Model Driven Architecture (General)Delphi Model Driven Architecture (Third Party)Delphi MultimediaDelphi Native APIDelphi Non-TechnicalDelphi ObjectPascalDelphi OLE AutomationDelphi OO designDelphi OpenToolsAPIDelphi Reporting-ChartingDelphi RTLDelphi StudentsDelphi TeamSourceDelphi Thirdparty ToolsDelphi Thirdparty Tools (General)Delphi Thirdparty Tools (IntraWeb)Delphi Thirdparty Tools (RaveReports)Delphi UpgradeDelphi VCLDelphi VCL ComponentsDelphi VCL Components UsingDelphi VCL Components WritingDelphi WebServices SOAPDelphi WinAPIDelphi XMLDelphi Installationcomp.lang.pascal.delphi.components.misccomp.lang.pascal.delphi.components.usagecomp.lang.pascal.delphi.components.writingcomp.lang.pascal.delphi.databasescomp.lang.pascal.delphi.misccomp.sources.delphi JBuilder Newsgroups----------------JBuilderJBuilder AnnouncementsJBuilder Applet IssuesJBuilder As400JBuilder CompilerJBuilder CORBA/RMIJBuilder DatabasesJBuilder DB SwingJBuilder DebuggerJBuilder DeploymentJBuilder DocumentationJBuilder EnterpriseJBuilder HandheldJBuilder IDEJBuilder JavaJBuilder JavaAPIJBuilder Java LanguageJBuilder Java SwingJBuilder Java BeansJBuilder Java Beans UsageJBuilder Java Beans DevelopmentJBuilder JBCLJBuilder JDataStoreJBuilder JobsJBuilder LinuxJBuilder MacintoshJBuilder & MobileSetJBuilder MultiLingual AppsJBuilder Non-TechnicalJBuilder OpenToolsJBuilder Servlets (JSP)JBuilder StudentsJBuilder Team DevelopmentJBuilder Thirdparty ToolsJBuilder UpgradeJBuilder Installation InterBase Newsgroups----------------InterBaseInterBase Bug DiscussionInterBase GeneralInterBase IBConsoleInterBase OpenSourceInterBase ReplicationInterBase SQLInterBase ThirdpartyInterBase Installation Miscellaneous Borland Newsgroups----------------AppCenterAppServerEnteraVisiBroker  You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum Powered by phpBB © 2001, 2006 phpBB Group SEO toolkit © 2004-2006 webmedic.