Hi, I would like to ask about the operation of code coverage instrumentation. When I performed instrumentation on the following code, the resulting instrumented code was as follows. 1) main.tcl puts "in main" source [file join [file dirname [info script]] proc.tcl] if {1} { myproc1 } else { myproc2 } 2) main.tcl_i ... # Initialise list of lines namespace eval ::instrument { set current test.dir/source.test.dir/main.tcl set idir {} set logFile main.tcl_log set log(test.dir/source.test.dir/main.tcl,6)...
Sorry... premature post... Cant' figure out how to delete.
Bumped revision to 1.3.5
More test for syntax comment
Corrected handling of syntax comment
Release fixes
Bumped revision to 1.3.4
Handle expression comments.
Small fix to syntax definition for encoding command in 9.0
I have made the same thing through syntaxbuild.tcl
Filling up with 8.7 and 9.0 additions
Added token "k".
Small fix to syntax definition for encoding command in 9.0
Change the error/warning message format
Add trace
The main problem is the removal of :: in parseStatement, which makes it impossible to distinguish the call to "::cmd" and "cmd" downstream in the code. I do not recall why it was added in 73aeb. But it all boils down to that the code mostly assumes "cmd" is global and that namespaces has been added later and not in a very good way. I think the namespace resolution needs to be moved to early in parseStatement, but it is not obvious at a glance how.
Merge /u/tbsmark86/nagelfar/ branch namespace-lookup into master
Fix unknown comands in nested namespaces
Some fixes & packagedb
Looks like this is a regression caused by commit 73aeb9166e75e33cd50bd4fd6d1c0c4a729458d2
False postive with fully qualified command names
Fix unknown comands in nested namespaces
Some fixes & packagedb
Tcl9 work
coverage issue with multile file soucing
Nagelfar binaries for Windows
Raw command substitution plugin hook
dir variable flagged as unknown for pkgIndex files
Fix committed
Handle pkgIndex.tcl specially. [Bug #43]
list element in quotes followed by "]" instead of space
Fix committed
Handle weird foreach lists nicely. [Bug #44]
Namespace and argument handling.
Clean up doc generation. Rerun with sphinx 5.3.0
list element in quotes followed by "]" instead of space
Tcl9 work
Preparing for Tcl 9. Dumping 8.5 support.
Merge /u/xdelaruelle/nagelfar/ branch codespell into master
Fix typos found with codespell
Fix copyright year range
dir variable flagged as unknown for pkgIndex files
Fix typos found with codespell
Inline comment documentation
textsearch lib licence mismatch
Fix copyright year range
Wonderful, thank you!
Raw command substitution plugin hook
I have made a release with this.
Bumped revision to 1.3.3
Updated syntax database
I just gave it a try and it is exactly what I need, thank you so much! I believe something like this is all I need now to perform the checks that would have been a real pain previously: proc bodyRaw {stmt info} { if {[dict get $info subst] && [string trim $stmt] ne $stmt} { lappend res error lappend res "Unneeded whitespace around '\['/'\]'s of command substitution" } } Out of curiosity, are you planning to make a new release that would include the new hook in the near future? No rush, but if it'll...
I just gave it a try and it is exactly what I need, thank you so much! I believe something like this is all I need now to perform the checks that would have been a real pain previously: proc bodyRaw {stmt info} { if {[dict get $info subst] && [string trim $stmt] ne $stmt} { lappend res error lappend res "Unneeded whitespace around '['/']'s of command substitution" } } Out of curiosity, are you planning to make a new release that would include the new hook in the near future? No rush, but if it'll...
I have committed a "bodyRaw" hook. Can you try it out?
New plugin hook bodyRaw
Hi Peter, thanks for your response. I would like the [] to be included in the call because part of our style guide mandates that command substitutions should have no whitespace between the opening brace and the first word of the command and the last word of the command and the closing brace, i.e. "[ foo]", "[foo ]", and "[ foo ]" are all in violation of our style guide. We expect those command substitutions to instead be written as "[foo]". If the [] are not included in the call, and the plugin only...
What you want is not readily available in an easy way. Adding a plugin hook there feels like a reasonable way. Why would you like the [] to be included in the call? That does not really add any info and does not really fit with how other hooks work. I also think if such a hook should really be a rawBody hook with an info-flag if the body is a command substitution body.
Raw command substitution plugin hook
Fix implicit variable typo on code and doc
Fix implicit variable typo on code and doc
Emit lint suggestions if there are similar commands
Merge accepted and polished a bit by me.
Polish unknown command plugin
The overall solution looks ok, so I have accepted this. I will look closer later and see if there is any polish needed.
Add Plugin Hook for unknown commands
Add Plugin Hook for unknown commands
Wish you a nice holiday! I got a prototype working and will send a merge request where we can discuss details further.
Quick answer since I'm on vacation. This sounds like a good idea, and my gut feeling is that this belongs in a plugin hook.
Emit lint suggestions if there are similar commands
Add support for raw line checks
Sorry that I have forgotten to look at this. I have rebased this to master and made a few fixes to make it run. I will look closer on the details later but I agree to include this functionality.
Rebuild databases
A bit dokumentation about database
Fix RawLine
Build updates
Add support for raw line checks
Update GPLv2 file and headers
Change URL in GUI to nagelfar.sourceforge.net
Update GPLv2 file and headers
Change URL in GUI to nagelfar.sourceforge.net
Track output
Setup local lexer for syntax highlighting