aboutsummaryrefslogtreecommitdiffhomepage
path: root/toplevel
Commit message (Collapse)AuthorAge
* Remove some uses of local modules (some were unused, some were costly).Gravatar xclerc2013-10-14
| | | | git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16883 85f007b7-540e-0410-9357-904b9bb8a0f7
* STM: prefix debug messages with slave-idGravatar gareuselesinge2013-10-11
| | | | git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16880 85f007b7-540e-0410-9357-904b9bb8a0f7
* More comments in ide_slaveGravatar gareuselesinge2013-10-11
| | | | git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16879 85f007b7-540e-0410-9357-904b9bb8a0f7
* STM: cancel slaves working on outdated jobsGravatar gareuselesinge2013-10-11
| | | | | | | | | | | | | | | | | | | | | | | | | I did not manage to make the slave manager use Unix.select to wait for a response from the slave for a limited time and check for cancellation. Hence the following semi-cooperative model: - The slave has a thread that sends a Tick every second when the thread is working - The slave_manager will then be unblocked periodically by this tick and check for cancellation - Cancellation is, for the moment, implemented using kill. To kill a process on windows one could bind TerminateProcess (>= WinXP) or RegisterWindowMessage + BroadcastSystemMessage (>= Win2k). See: http://msdn.microsoft.com/en-us/library/windows/desktop/ms686722%28v=vs.85%29.aspx Another option is to make the slave_manager send to the tick thread on the slave process a boolean answer to the Tick message, and the tick thread could eventually bail out. But to do that, it is way better to have a second channel, used only by this tick thread. This solution sound very like the one proposed for windows, but requires more work. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16878 85f007b7-540e-0410-9357-904b9bb8a0f7
* STM: a proof with nested proofs cannot be delegatedGravatar gareuselesinge2013-10-10
| | | | | | | | | The reason is that the state gets altered by side effects by the Qed of inner proofs. This kind of side effects cannot be reproduced in the slaves easily. And there is no point in working hard for this corner case. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16869 85f007b7-540e-0410-9357-904b9bb8a0f7
* STM: add "Stm Wait" to wait for the slaves to complete their jobsGravatar gareuselesinge2013-10-10
| | | | | | | Used by fake_ide, that before editing a broken proof has to be sure Coq known the proof is broken. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16868 85f007b7-540e-0410-9357-904b9bb8a0f7
* Clib: fold_left_until added to CListGravatar gareuselesinge2013-10-10
| | | | | | | CStack just calls it to implement fold_until. CSig.seek renamed CSig.until, since there is no seek function. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16867 85f007b7-540e-0410-9357-904b9bb8a0f7
* STM: new command "Stm PrintDag" to force printing the dag to /tmpGravatar gareuselesinge2013-10-07
| | | | git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16862 85f007b7-540e-0410-9357-904b9bb8a0f7
* STM: fix verbosity of queriesGravatar gareuselesinge2013-10-07
| | | | git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16861 85f007b7-540e-0410-9357-904b9bb8a0f7
* STM: spit a warning if an out of bound Back* command is issuedGravatar gareuselesinge2013-10-07
| | | | git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16859 85f007b7-540e-0410-9357-904b9bb8a0f7
* coqtop: init STM before loading rcfileGravatar gareuselesinge2013-10-07
| | | | git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16858 85f007b7-540e-0410-9357-904b9bb8a0f7
* Removing uses of Evar.add in class-related functions.Gravatar ppedrot2013-10-06
| | | | git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16852 85f007b7-540e-0410-9357-904b9bb8a0f7
* Removing dubious use of evarmap manipulating functions in printingGravatar ppedrot2013-10-05
| | | | | | related code. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16851 85f007b7-540e-0410-9357-904b9bb8a0f7
* Moving side effects into evar_map. There was no reason to keep anotherGravatar ppedrot2013-10-05
| | | | | | | | | | | | | | | | | | | state out of one we were threading all the way along. This should be safer, as one cannot forego side effects accidentally by manipulating explicitly the [sigma] container. Still, this patch raised the issue of badly used evar maps. There is an ad-hoc workaround (i.e. a hack) in Rewrite to handle the fact it uses evar maps in an unorthodox way. Likewise, that mean we have to revert all contrib patches that added effect threading... There was also a dubious use of side effects in their toplevel handling, that duplicates them, leading to the need of a rather unsafe List.uniquize afterwards. It should be investigaged. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16850 85f007b7-540e-0410-9357-904b9bb8a0f7
* STM: understand -coq-slaves-opts extra-env=VAR=valGravatar gareuselesinge2013-10-03
| | | | | | | In this way one can set an env variable for the slave, like the socket used by ocamldebug to talk remotely to a process git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16844 85f007b7-540e-0410-9357-904b9bb8a0f7
* STM: if -coq-slaves off really imitate the old CoqIDEGravatar gareuselesinge2013-10-03
| | | | git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16843 85f007b7-540e-0410-9357-904b9bb8a0f7
* STM: add options to disable fallbacks to ease regression testingGravatar gareuselesinge2013-10-03
| | | | | | | | | | STM falls back to local, lazy, evaluation if the slave dies badly or if there is a marshalling error. Both things should never occur, but is nice to have the system recover if a bug pops up. Nevertheless during regression testing these fallbacks should be disabled not to hide a new bug. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16841 85f007b7-540e-0410-9357-904b9bb8a0f7
* Regression test suite for STMGravatar gareuselesinge2013-10-03
| | | | git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16840 85f007b7-540e-0410-9357-904b9bb8a0f7
* STM: number of slaves passed by the command lineGravatar gareuselesinge2013-10-03
| | | | git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16839 85f007b7-540e-0410-9357-904b9bb8a0f7
* STM: avoid "kill 9 pid" if we know the slave diedGravatar gareuselesinge2013-10-03
| | | | git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16838 85f007b7-540e-0410-9357-904b9bb8a0f7
* STM: delegate proofs to slaves only if they are not trivialGravatar gareuselesinge2013-10-03
| | | | | | Still too simple and not configurable git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16837 85f007b7-540e-0410-9357-904b9bb8a0f7
* STM: take a shallow snapshot for the first proof stepGravatar gareuselesinge2013-10-03
| | | | | | this way it can be directly sent to a slave with no further manipulation git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16836 85f007b7-540e-0410-9357-904b9bb8a0f7
* STM: fully force a proof ourput before sending it back to the masterGravatar gareuselesinge2013-10-01
| | | | git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16830 85f007b7-540e-0410-9357-904b9bb8a0f7
* STM: better error messages in case of marshal failureGravatar gareuselesinge2013-10-01
| | | | git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16829 85f007b7-540e-0410-9357-904b9bb8a0f7
* STM: fix reporting of ongoing tasksGravatar gareuselesinge2013-10-01
| | | | git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16828 85f007b7-540e-0410-9357-904b9bb8a0f7
* STM: exceptions caming from slaves are now localizedGravatar gareuselesinge2013-10-01
| | | | git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16827 85f007b7-540e-0410-9357-904b9bb8a0f7
* STM: Unix.kill can failGravatar gareuselesinge2013-10-01
| | | | git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16825 85f007b7-540e-0410-9357-904b9bb8a0f7
* STM: better handle proof modesGravatar gareuselesinge2013-09-30
| | | | | | | | Proof modes are really spaghetti code. It is a global state that you can't access (held by G_vernac). We stick it to the branches... git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16820 85f007b7-540e-0410-9357-904b9bb8a0f7
* STM: some refactoring, support revised CoqIDE protocolGravatar gareuselesinge2013-09-30
| | | | git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16815 85f007b7-540e-0410-9357-904b9bb8a0f7
* CoqIDE ported to the revides protocolGravatar gareuselesinge2013-09-30
| | | | | | | | | | | - the zone to be edited is now between the marks start_of_input and stop_of_input - when -debug is given, the edit zone is underlined - the cmd_stack is focused/unfocused according to the new protocol - read only tag resurrected and used to block the "Qed" ending a focused zone git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16814 85f007b7-540e-0410-9357-904b9bb8a0f7
* Removing a bunch of generic equalities.Gravatar ppedrot2013-09-27
| | | | git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16806 85f007b7-540e-0410-9357-904b9bb8a0f7
* Made the filename of compiled files explicit, i.e. add a ./ prefixGravatar ppedrot2013-09-19
| | | | | | | | | | | whenever the filename was given alone. This caused strange behaviour of coqtop that was compiling stuff we didn't want to just because it was somewhere in its loadpath. This is why FingerTree was not compiling, btw, as it had two files equally named in two different paths. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16788 85f007b7-540e-0410-9357-904b9bb8a0f7
* Get rid of the uses of deprecated OCaml elements (still remaining compatible ↵Gravatar xclerc2013-09-19
| | | | | | with OCaml 3.12.1). git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16787 85f007b7-540e-0410-9357-904b9bb8a0f7
* At least made the evar type opaque! There are still 5 remaining unsafeGravatar ppedrot2013-09-18
| | | | | | | | | | | | casts of ints to evars. - 2 in Evarutil and Goal which are really needed, even though the Goal one could (and should) be removed; - 2 in G_xml and Detyping that are there for completeness sake, but that might be made anomalies altogether; - 1 in Newring which is quite dubious at best, and should be fixed. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16786 85f007b7-540e-0410-9357-904b9bb8a0f7
* Removing almost all new_untyped_evar, and a bunch of Evd.add.Gravatar ppedrot2013-09-18
| | | | | | | Ultimately all evars should be created with respect to a given evar map, instead of using a global counter. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16783 85f007b7-540e-0410-9357-904b9bb8a0f7
* STM pretty printer should never failGravatar gareuselesinge2013-09-13
| | | | git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16777 85f007b7-540e-0410-9357-904b9bb8a0f7
* fix STM feeback on running jobsGravatar gareuselesinge2013-09-13
| | | | git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16776 85f007b7-540e-0410-9357-904b9bb8a0f7
* Unplugging Autoinstance. The code is still here if someone wishesGravatar ppedrot2013-09-12
| | | | | | to fix it, but it should be eventually erased. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16774 85f007b7-540e-0410-9357-904b9bb8a0f7
* CoqIDE: show number of proofs being checked in backgroundGravatar gareuselesinge2013-09-12
| | | | | | good test: Nijmegen/QArithSternBrocot/Zaux.v git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16773 85f007b7-540e-0410-9357-904b9bb8a0f7
* Unknown commands starting a proof were not setting the proof mode.Gravatar gareuselesinge2013-09-12
| | | | | | Fixes Orsay/Containers git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16771 85f007b7-540e-0410-9357-904b9bb8a0f7
* Remove outdated commentGravatar gareuselesinge2013-09-12
| | | | git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16770 85f007b7-540e-0410-9357-904b9bb8a0f7
* VernacList handling was lost in STM mergeGravatar gareuselesinge2013-09-12
| | | | git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16769 85f007b7-540e-0410-9357-904b9bb8a0f7
* Temporary fix of emacs mode for ProofGeneralGravatar letouzey2013-09-10
| | | | | | | | | | To the best of my knowledge, even with the latest cvs version of PG: - the goals must be finally displayed at each step - the <info>...</info> async feedback messages aren't handled yet. I'll check with Enrico what he intended with these two lines, but meanwhile let's comment them... git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16767 85f007b7-540e-0410-9357-904b9bb8a0f7
* Moving Searchstack to CStack, and normalizing names a bit.Gravatar ppedrot2013-09-06
| | | | git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16765 85f007b7-540e-0410-9357-904b9bb8a0f7
* Partly replacing list-based access functions in Evd. This is stillGravatar ppedrot2013-09-03
| | | | | | | unsatisfactory as some functions implicitly require some ordering on the evars, but this is already better. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16759 85f007b7-540e-0410-9357-904b9bb8a0f7
* summary for ML modules made correctGravatar gareuselesinge2013-08-30
| | | | | | | | | This summary entry is special because its unfreeze may load ML code that in turns adds summary entries. Hence it is the first summary piece to be unfreezed, otherwise some summaries may get lost. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16752 85f007b7-540e-0410-9357-904b9bb8a0f7
* Make the jopin-document button wait for slaves to terminateGravatar gareuselesinge2013-08-30
| | | | git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16750 85f007b7-540e-0410-9357-904b9bb8a0f7
* When PG is used as interface behave as before STMGravatar gareuselesinge2013-08-30
| | | | git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16748 85f007b7-540e-0410-9357-904b9bb8a0f7
* ind_tables: properly handling side effectsGravatar gareuselesinge2013-08-30
| | | | | | | | | | | | | | If a constant is defined as transparent, not only its side effects (opaque sub proofs as in abstract, and transparent ind schemes) are declared globally, but the ones that are schemes are also declared as such. The only sub optimal thing is that the code handling in a special way the side effects of transparent constants is in declare.ml that does not see ind_tables.ml, hence a forward ref to a function is used. IMO, ind_tables has no reason to stay in toplevel/. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16747 85f007b7-540e-0410-9357-904b9bb8a0f7
* Stm: if slave process dies badly go back to local lazy evaluationGravatar gareuselesinge2013-08-30
| | | | | | | | | | | Processing the proof in a slave process may fail for an implementation error, e.g. the state could not be marshalled, or an anomaly is raised by the slave. In this case we fall back to local, lazy, evaluation in the master process. git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@16743 85f007b7-540e-0410-9357-904b9bb8a0f7
generated by cgit v1.2.3 (git 2.39.1) at 2024-06-29 21:45:09 +0000