TPP:V5Feedback

TPP 5.0 pre-release tasks

• Update the TPP wiki tutorial
• Update the TPP wiki installation instructions
• Update the TPP wiki build instructions
• I'd recommend finishing the new wiki/replace old wiki project
  • go through the current pages and update/remove them

Feedback related to TPP version 5.0 builds

This page here is to record feedback/issues found with the development builds of 5.0.0

Release 5.0.0a5

• nothing yet

Release 5.0.0a4

• Uninstall reports error when tpptray is running and in system tray. Should shut down executable before deleting
  • Joe: only an issue with TPP's "make uninstall". Works as expected when run from command shell or add/remove programs.

• Possible bug with the tpp data directory. When you have install over TPP 4.8, it sets the default tpp data directory to the old directory under Inetpub. But I don't think it's updating it everywhere as it seemed the web interface was looking in the new area.
  • Joe: changed priority of default TPP_DATA environment variable, if that doesn't exist then old ISB directory, if that isn't there then {app}/data.

• Launching TPP Tray pops up a command window, despite using wperl.exe. Largely because we associate .pl with perl.exe instead of wperl.exe. Not a big deal as TPP tray should be launched at startup when you log in. Could turn this into a .bat file and get rid of this problem.

• Like to include tpp tray in the startup menu but that then screws up the order.

• On my system seems to keep trying to install the VS 2005 runtime every install.

• BUILD_LINUX instructions are way out of date. Also take a look at the BUILD_OSX
  • mention (or otherwise notify users) that make install only works in top directory because of dependencies

• Interesting issue -- when shutting down appweb, any jobs continue to run. Can't restart appweb (port 80 in use) until jobs finish or are killed in taskmanager

• Any call to gnuplot.exe results in "Fontconfig error: Cannot load default config file"

• Seems to be a limit to post/put forms found by deleting 25+ files in petunia
  • Joe: doesn't seem to be a limit as I tested by setting available directives to high settings. Also seen situation where it seems to forget the login session
  • Joe: reported as an appweb bug: https://github.com/embedthis/appweb/issues/530. They say they won't fix it.

• legacy proteinprophet results: prot_wt_xml.pl: "The system cannot find the path specified."
  • ASAPRatio link to ASAPCGIDisplay.cgi on the resulting page crashes
  • pvalue link to interact-pval.png is broken. (look in PValueModel.cxx)

• ASAPRatioPeptideCGIDisplayParser.cgi, call to unsupported function: rampValidateOrDeriveInputFilename(char *inbuf, int inbuflen, char *spectrumName)
  • Also reports 'gnuplot.exe' is not recognized as an internal or external command, operable program or batch file.

• Clicking on the ions2 link to view the spectrum in comet results in the pepxmlviewer results in a bad url error

• Clicking on the ions link in the comet pepxml view of the top 5 matching peptides results in a bad url error

• Clicking on the 'st' in the pepxmlviewer to view the spectrast results at peptideatlas is broken

Release 5.0.0a3

• Haven't merged DDS/Luis changes from today in.

• New batch of DDS changes have to be merged

• PWiz Window's runtime dependencies can fail to install causing TPP installer to fail. Other than not installing pwiz no way to get past this.
  • Joe: Made it an option

• comet-fastadb.cgi crashes if database doesn't exist (or for that matter if no cgi options are provided).

• multi_parm() calls in protxmlviewer

• Links are bad in pep3d save as html document. Resulting page has a bad link (windows path) to the image. Opening the html file also doesn't link to image

• Export spreadsheet from pepxmlviewer doesn't seem to do anything Joe: nevermind seems to work

• Extra '/' in link for Additional Analysis Info on pepxmlviewer

• Can't remove mzML files from processing search form

• Clicking on spectrum id in the pepxmlviewer when viewing results of a spectrast search result in:
  • command "gnuplot.exe c:/TPP/data/spectrast/OR20080317_S_SILAC-LH_1-1_01.00417.00417.2.gp" failed: Operation not permitted command "gnuplot.exe c:/TPP/data/spectrast/OR20080317_S_SILAC-LH_1-1_01.00417.00417.2.gp" failed with exit code 1 'gnuplot.exe' is not recognized as an internal or external command, operable program or batch file.
  • URL is plotspectrast.cgi?LibFile=c:/TPP/data/dbase/speclibs/NIST_yeast_IT_v2.0_2008-07-11.splib&LibFileOffset=413954123&QueryFile=c:/TPP/data/spectrast/OR20080317_S_SILAC-LH_1-1_01.mzML&QueryScanNum=00417

• Lumping these all together under "protein identifiers don't seem to be parsed correctly. Contain '|' NCBI field seperators"
  • Clicking on the protein link in the pepxmlviewer when viewing spectrast results causes an cgi crash with "400 - Bad Request, Bad URL"
    • url: comet-fastadb.cgi?Ref=gi|6321631|ref|NP_011708.1|&Db=c:/TPP/data/yeast_orfs_all_REV.20060126.short.fasta&Pep=YDSTHGR&MassType=1&sample_enzyme=trypsin&min_ntt=0
    • It really looks like the Ref isn't escaped. Clicking on a ref with a plain text string works
  • Clicking on the protein link in the pepxmlviewer when viewing peptideprophet results of spectrast results causes an cgi crash with "400 - Bad Request, Bad URL" (see above)
  • Joe: now urlencode the cgi parameters so it won't crash. But it also doesn't find the protein
  • Joe: update -- luis says this is the way its always worked

• Issue with extracted ion current, launched from pepxmlviewer on spectrast results
  • May be that d3.js is not found "localhost/ISB//html/js/d3.js"
  • Joe: fixed multiple issues

• Same gnuplot error as before, on pepxmlviewer when viewing peptideprophet results of spectrast results. Displayed when selecting the spectrum inthe pepxmlviewer on spectrast results.

• plotspectrast.cgi has bad image link in it (http://maserati/spectrast/OR20080317_S_SILAC-LH_1-1_01.00417.00417.2.png) Seen when clicking on a spectrum in the pepxmlviewer on spectrast results.

• ProteinProphet spits this error:
  • 'gnuplot.exe' is not recognized as an internal or external command, operable program or batch file.

• proteinprophet results: new viewer doesn't display anything in pepxmlviewer when a peptide is clicked on (no peptides shown in view)

• proteinprophet results: saving to JSON in new viewer reported error "Service Unavailable"

• legacy proteinprophet results: sensitivity/error info has broken image/page with error "The system cannot find the path specified."

• legacy proteinprophet results: more analysis info reports error "The system cannot find the path specified."

• legacy proteinprophet results: help link has page that's missing images

Release 5.0.0a2

• Uninstall leaves cmd.exe, and data/ log/ users/ dirs. (by design?)
  • Joe: Yes. Well at least data/, log/. Don't think you'd want to remove TPP and lose all your data. The users/ directory could get removed, and I guess the log directory if you guys think so. The cmd.exe file, is that in bin? If so it should have been removed.
  • Joe: Just retested and it only left data/ behind.
• ERROR in Petunia. CGI::multi_param is not a valid subroutine (new version of CGI.pm)
  • Joe: This again?!?! So if you use ActivePerl you'll need to update CGI, I sent along instructions in a email to Luis. Otherwise yes, there still are places where parm() is called in a array context and needs to be replaced with multi_param(). I did another pass looking for these occurrences, but there are many. Someone other than me should do a grep 'param(' and check them out.
• Adjust web timeout!?? Seemed to have pooped at 2 minutes...? FIXED. No longer takes a poop break.
• PepXMLViewer: BAD link to SpectraST ("ST") FIXED (I think. Is it now doing what its suppose to)
• Petunia icon on desktop
  • Window's quirk? Sometimes I see this sometimes I don't. The code in the installer looks solid. FIXED
• Petunia favico in browser tab is missing on some pages
  • Joe: as far as I can tell, Petunia is the only one that actually supplies a favico. Otherwise it defaults to what the browser provides. On regis its that regis icon. I don't know what it is on windows. I looked but didn't see where we might have setup an favico in the apache config file. Can you check on a older TPP windows installation? Oh, and I didn't see a way to configure one for Appweb in the documentation however if you do place a favicon.ico file at the document level it seems to work, though your browser caching may hide this fact.
  • Joe: also, keep in mind that the browser will stubbornly cache the icon and display it. http://stackoverflow.com/questions/2208933/how-do-i-force-a-favicon-refresh
• tpp_models: bad links to js/ resources ("/ISB/html...") FIXED
• missing jquery.flot.errorbars.js. FIXED
  • Joe: as far as I can tell this was never a part of TPP. However the file was being copied into the TPP installation on regis by lmendoza. So I added it to the distribution.
• ProtXMLViewer fails...(bad gateway) 'FIXED;
• Old ProtXMLViewer fails too: system cannot find the path specified FIXED
• updatepaths does not seem to do anything, no output, even with -vv FIXED
• mzIdent conversion fails Potentially fixed with the perl script fix?
• Petunia::Issue when selecting a directory from quick-select drop-down
  • Joe: not sure what this means.
• decoyFastaGenerator.pl does not launch from Petunia FIXED
• Downloading Spectral Libs fails, yet again: bad paths to md5sum, wget, and unzip FIXED

Release 5.0.0a1

• Merge last set of trunk changes into branch

• Review one last time what's in external. Ditch unneeded stuff.

• Add instructions on wiki on how to use your own Perl/Apache
  • (Joe) warn users if they choose not to install perl or appweb Apache

• Luis says the TPP icon is ugly and use the purple flower instead

• Update linux scripts in installer_linux

• Update BUILD_LINUX. Especially Apache instructions

• Luis says msconvert fails.
  • Luis is right<\s>