fche changed the topic of #systemtap to: http://sourceware.org/systemtap; email systemtap@sourceware.org if answers here not timely, conversations may be logged
<fche> yeah that looks good
<agentzh> cool, i'll prepare a formal patch with tests then.
<agentzh> fche: btw, it seems that stap does not support the al register?
<agentzh> (of x86)
<agentzh> we could add that?
<agentzh> though we could always use bit op on ax or eax, but builtin support for al/bh/etc can still be handy :)
<fche> no strong opinion on that
<agentzh> hmm, seems like the x86_64 register tapset func does not really support 16-bit registers it seems. it always reads 32-bit from those 16-bit registers like ax.
<agentzh> is it deserved fixing?
* agentzh is using stap to probe assembly programs.
orivej has joined #systemtap
slowfranklin has joined #systemtap
slowfranklin has quit [Client Quit]
orivej has quit [Ping timeout: 246 seconds]
orivej has joined #systemtap
naveen4 has quit [Quit: WeeChat 1.9.1]
naveen has joined #systemtap
orivej has quit [Ping timeout: 252 seconds]
mjw has joined #systemtap
mjw has quit [Quit: Leaving]
slowfranklin has joined #systemtap
mjw has joined #systemtap
orivej has joined #systemtap
pwithnall has joined #systemtap
ericlee has quit [Remote host closed the connection]
orivej has quit [Ping timeout: 245 seconds]
orivej has joined #systemtap
brolley has joined #systemtap
orivej has quit [Ping timeout: 245 seconds]
orivej has joined #systemtap
tromey has joined #systemtap
orivej has quit [Ping timeout: 252 seconds]
mjw has quit [Quit: Leaving]
slowfranklin has quit [Quit: slowfranklin]
pwithnall has quit [Read error: Connection reset by peer]
pwithnall has joined #systemtap
pwithnall has quit [Ping timeout: 252 seconds]
slowfranklin has joined #systemtap
slowfranklin has quit [Quit: slowfranklin]
slowfranklin has joined #systemtap
slowfranklin has quit [Quit: slowfranklin]
slowfranklin has joined #systemtap
slowfranklin has quit [Ping timeout: 245 seconds]
<agentzh> is there a web page detailing how to subscribe to the systemtap mailing list? googling does not reveal anything helpful, it seems.
<fche> https://sourceware.org/lists.html ... or the headers of the emails
<agentzh> okay, thanks.
<agentzh> seems like this page is better: https://sourceware.org/systemtap/getinvolved.html
<fche> that too
<fche> it's also mirrored to nntp at gmane.org if you roll that way
slowfranklin has joined #systemtap
slowfranklin has quit [Client Quit]
<agentzh> got it.
orivej has joined #systemtap
<agentzh> fche: abort patch committed with the "flat" typo fix.
<fche> thanks
<fche> and thanks for your patience!
<agentzh> sure thing, thanks for your quick response! :)
slowfranklin has joined #systemtap
slowfranklin has quit [Quit: slowfranklin]
brolley has left #systemtap [#systemtap]
slowfranklin has joined #systemtap
orivej has quit [Ping timeout: 245 seconds]
slowfranklin has quit [Quit: slowfranklin]
tromey has quit [Quit: ERC (IRC client for Emacs 26.1.50)]
orivej has joined #systemtap
<agentzh> fche: just sent the bare return v4 patch: https://sourceware.org/ml/systemtap/2018-q3/msg00150.html
<fche> looks good
<fche> just missing the part in the exp file where e.g. no_val_11.stp is processed
orivej has quit [Ping timeout: 240 seconds]
<agentzh> fche: oh, right, i forgot to remove those obsolete *.stp files. will fix that.
<agentzh> thanks for the catch.
<fche> wouldn't say necessarily obsolete
<fche> 11 checks an interesting case, should be a parse error
<agentzh> right, i'll add one.
<agentzh> fche: how about this extra diff? https://pastebin.com/JtqYEj32
<agentzh> i re-added the original tests but to check syntax errors.
<fche> ok
<fche> ya know, there is a dejagnu library proc for running a stap script and comparing its output
<fche> so maybe your test cases don't have to be quite so wordy
<agentzh> yeah, we could add a proc to testsuite/lib/ to abstract certain common parts.
<agentzh> it's just that it's the most flexible way for my test generator :)
<agentzh> so it can now be committed? ;)
<fche> sure
<agentzh> awesome.
<agentzh> fche: do you mind adding a resusable proc to testsuite/lib? if not, i can prepare a patch.
<fche> there may already be one
<fche> the stap_run* series
<agentzh> i was aware of those but unfortunately they do not work for my purposes.
<agentzh> i need to test stdout, stderr, and exit code separately.
<fche> sure, a new proc then
<fche> the less boilerplate in our files, the better
<agentzh> i understand.
<agentzh> should i name it stap_run4?
<fche> if you like
<agentzh> it's a horrible name though.
<fche> make it a better name
<agentzh> i'll think about it.
<agentzh> after having the new proc, i'll update all my tests to use it.
<agentzh> they were generated by my tool anyway.
<fche> yup
<agentzh> how about [stap_run_ok cmd exp_out exp_out_is_regex exp_stderr exp_stderr_is_regex exp_exit_code] ?
<agentzh> the exp_out_is_regex and exp_stderr_is_regex are boolean values.
<fche> stap_run_regexps ?
<agentzh> oh, sorry, exp_exit_code is not needed for stap_run_ok.
<agentzh> if we could also have a stap_run_nok proc
<agentzh> or maybe we could simply check whether the exit code is zero or not.
<agentzh> none of my existing tests actually check the non-zero exit code's actual values anyway.
<agentzh> the problem with stap_run_regex is that we could have literal exp_stdout but regex exp_stderr (or the other way around)
<agentzh> or we could introduce several procs for different combinations just like stap runtime's map_set_is/map_set_ss and etc?
<agentzh> like run_stap_rs_ok or run_stap_sr_ok?
<agentzh> or run_stap_rr_ok
<fche> rs?
<agentzh> rs == regex string
<agentzh> or rl == regex literal if you like it better...
<agentzh> personally i like separate boolean args better.
<agentzh> it's easier to remember
<agentzh> and easier to implement.
<fche> or 'regex' :)
<agentzh> 'regex' for which part?
<fche> proc name
<fche> no need to abbreviate so much that it becomes a mystery
<agentzh> you mean something like run_stap_ok_regex_str ?
<agentzh> or even run_stap_ok_rx_str ?
<agentzh> rx is a common regex abbrev i think.
<agentzh> or rx_lit
<agentzh> for regex stdout pattern and literal stderr pattern
<agentzh> similarly, we can have run_stap_nok_lit_rx and etc.
<agentzh> nok is a bit confusing.
<agentzh> it could mean the stdout pattern should not match...hmmm
<agentzh> not good.
<agentzh> or maybe we should provide separate pros for running stap and checking stdout/stderr/exit code?
<agentzh> so that we do not have squeeze all the combinations into a single poor proc :)
<agentzh> *have to
<fche> your choice, think about it
<fche> don't feel pressured to make the proc names super compact
<fche> and don't feel pressured to make only one super parametrized proc, if more specialized procs make sense
<agentzh> just [run_stap4 stdin stdout stderr exit_code] and then [like stderr stderr_regex "$desc"] [is stdout exp_stdout_literal "$desc"] ?
<agentzh> now run_stap4 is a nifty name since it outputs 4 things.
<agentzh> oh, i missed the input arg, cmd, as the first arg.
<agentzh> should be [run_stap4 cmd stdin stdout stderr exit_code]
<agentzh> the names "like" and "is" are from perl's testing world. if you think they are too short, maybe we could name them "like_ok", "is_ok", and etc.
<agentzh> but i really like "like", "unlike", "is", "isnt", "ok", and "nok" subs in the perl testing world.
<agentzh> very handy.
<agentzh> i'll work out a patch in the next couple of hours if you like this design.
* agentzh hates verbose code too.
fche has quit [Quit: Leaving]