Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
    Carl Anderson
    @MobileOak
    I'm trying to download and install serge on my Mac
    but the instructions aren't working
    wget isn't present as far as I can tell, so I used curl to get the serge-1.3.zip file
    but then when I try to unzip the file I get an error saying that it's not a zipfile
    Taras Semenenko
    @ta2-1
    Hi! you can install wget via brew.
    Carl Anderson
    @MobileOak
    Ok, but if I wget the file is it going to be different than just grabbing it with curl?
    huh.. apparently it is
    Igor Afanasyev
    @iafan
    @MobileOak curl doesn't do redirects. I recommend using the latest version. Try this command:
    curl https://codeload.github.com/evernote/serge/zip/master >serge-master.zip
    and then as the installation doc suggests:
    unzip serge-master.zip
    unlink serge-master.zip
    cd serge-master
    Carl Anderson
    @MobileOak
    Oh I just used brew to get wget and am good now
    Igor Afanasyev
    @iafan
    :+1:
    ekaterinag
    @ekaterinag
    Hello all. There is a problem when serge pull config.serge : "Aborting unfinished rebase
    HEAD is now at 83868aa Merged PR 97919: #3337458". In this case we delete git folder and run "serge pull --initialize config.serge"... I think it is wrong way... Is there another solution to avoid "Aborting unfinished rebase
    HEAD is now at 83868aa Merged PR 97919: #3337458"?
    Igor Afanasyev
    @iafan
    @ekaterinag this message comes from Git, which means and automatic rebase of the local checkout is not possible. The most likely cause of unrecoverable rebase errors is when the history in the upstream repo has changed. Was that the case? (And when there's an unrecoverable rebase error, the simpleset solution is the one you did: remove the existing local repo and initialise it with serge pull --initialize.)
    ekaterinag
    @ekaterinag
    Thank you ! But it is not good wY for automatic way
    Dragos Varovici
    @dragosv
    @iafan any progress with releasing serge 1.4?
    ekaterinag
    @ekaterinag
    Hello all! Is there any way to transform %PATH%: replace "/" to "." or smth else to be added to ts files as preffix? Ex, we have many files "./project/kit1/values/strings.xml, ./project/kit2/values/strings.xml,..., ./project/kitN/values/strings.xml"; the po files must be "project.kit1.values.strings.xml.po, ..." or "project.kit1.strings.xml.po". Callbacks plugins?
    Taras Semenenko
    @ta2-1
    Hi Katya! You can use this option https://serge.io/docs/dev/callbacks/#rewrite_path
    Though I’m going to implement this on the smartcat plugin side)
    ekaterinag
    @ekaterinag
    thanks :)
    JAIN K JOHNSON
    @JAINKJOHNSON_twitter

    Hi
    I integrated postgresql with serge and I got a error like this.

    Initializing the database...
    Exception occurred while processing job 'my_project.job': Can't open schema file /usr/lib/serge-master/bin/../lib/Serge/pg_schema.sql: No such file or directory

    Then I looked the github repo, and I didn't see pg_schema.sql but I can see schemas for mysql and sqlite3.
    Is there any way to tackle it.

    Thanks in advance :)

    Igor Afanasyev
    @iafan

    @JAINKJOHNSON_twitter potentially Serge can work with Postgress (as with any other RDBMS that has a DBI driver); practically, it was tested and currently supports only MySQL and SQLite. When you use any driver in the database connection string, and the database doesn't exist, Serge tries to locate the lib/Serge/xx_schema.sql file where xx is the driver name. For postgress that would be pg. And Serge doesn't ship with such a file (it only has mysql_schema.sql and sqlite_schema.sql). This shema file is a set of SQL statements that one needs to run on a blank database to initialize all the tables and indices. If you want to use Postgres, you can take mysql_schema.sqlfile as a starting point and create pg_schema.sql from it, possibly altering for Postgres-specific syntax. If that works, feel free to do a PR with the schema file.

    Having said that, I typically encourage everybody to start with an SQLite database. It's fast, simple, and usually more than enough for practical purposes. The database is a single file that you can easily backup and restore.

    JAIN K JOHNSON
    @JAINKJOHNSON_twitter
    @iafan Thanks for the reply. I will try to create a pg_schema.sql file from mysql_schema.sql.
    locpeople
    @locpeople
    Hi all! I hope you can help me understand some high-level concept regarding Serge, I'm thinking we could link it to our (yet to be acquired) XTM tenancy. My questions would be:
    1) would that make any sense at all?
    2) XTM has its own TM management and so does Serge. Should we use either one or the other, or both, or how would one generally approach this?
    Thanks
    Dragos Varovici
    @dragosv
    @locpeople yes, you can use both TM. This is because you can reuse translations between different projects in Serge.
    I've implemented plugins for a few TMS (https://github.com/dragosv/serge-plugins), however I have not covered XTM.
    I believe support for XTM is doable, but I am not seeing a command line interface for it, plus it may only be job based, meaning you have to send the changed strings to be translated, instead of sending all strings and the TMS figuring out the diff.
    locpeople
    @locpeople
    Thanks @dragosv. XTM doesn't have a CLI as far as I know but it does have a REST API and a "watched" directory to drop files into for localization. And yes, it is job based only
    Igor Afanasyev
    @iafan

    @locpeople even if XTM has an API, it won't give you a proper continuous localization workflow (because of the jobs mentioned above) and ultimately won't allow you to approach proper translation quality (all jobs are translated out of global context). So I'd strongly recommend finding some solution that is not job-based.

    Answering your questions on TM, Serge accumulates translations in its own database, but there's no 'management' part there. You will simply have translations stored in Serge in addition to whatever CAT frontend you use stores.

    .
    @elrasskazenok_twitter
    Hello, I have a problem:

    Test Summary Report

    t\engine.t (Wstat: 65280 Tests: 12 Failed: 1)
    Failed test: 12
    Non-zero exit status: 255
    Parse errors: No plan found in TAP output
    Files=2, Tests=116, 13 wallclock secs ( 0.23 usr + 0.02 sys = 0.25 CPU)
    Result: FAIL
    Failed 1/2 test programs. 1/116 subtests failed.
    -> FAIL Installing Serge failed. See C:\Users\User.cpanm\work\1560419535.9068\build.log for details. Retry with --force to force install it.
    -> FAIL Installing the dependencies failed: Missing version info for module 'Serge'
    -> FAIL Bailing out the installation for Serge-Sync-Plugin-TranslationService-locize-0.900.0.
    How can I resolve it?
    Igor Afanasyev
    @iafan

    @elrasskazenok_twitter I think it could be due to evernote/serge#85

    For now, since this is the only test failing, you can try installing this with the --force flag. Should work fine.

    Dragos Varovici
    @dragosv
    @iafan any progress with evernote/serge#89 ?
    Igor Afanasyev
    @iafan
    @dragosv planning to work on this one soon. Definitely want it to be a part of 1.4 release that I'm currently preparing.
    Timur Gilfanov
    @timurgilfanov
    Hi! I trying to build docker image on Mac with this command in root of serge repository:
    serge git:(master) docker build docker
    It ends with this line:
    "+ cpanm --no-wget --installdeps /serge
    --> Working on /serge
    Configuring /serge ... N/A
    ! Configuring /serge failed. See /root/.cpanm/work/1562233477.5178/build.log for details.
    The command '/bin/sh -c set -ex && echo ${INSTALL_PREREQUISITES} > /tmp/prereq && . /tmp/prereq && rm /tmp/prereq && cpanm --no-wget --installdeps /serge && rm -rf /root/.cpan* /usr/local/share/man && mkdir -p /data && cp /serge/doc/sample.serge /data && echo ${CLEAN_PREREQUISITES} > /tmp/prereq && . /tmp/prereq && rm /tmp/prereq' returned a non-zero code: 1"
    I'm new to docker, so maybe I build image with wrong command or don't do something important.
    Igor Afanasyev
    @iafan
    @timurgilfanov you need to run your semi-built docker image and examine the contents of the /root/.cpanm/work/1562233477.5178/build.log file that was left after the installation. It will contain the information on what Perl module failed to be installed. I'll try to reproduce this locally as well and let you know.
    Igor Afanasyev
    @iafan
    Ok, I was able to reproduce the error; will investigate further
    Timur Gilfanov
    @timurgilfanov
    @iafan "docker build -f docker/Dockerfile ." solves the problem. I told you i'am new to docker :)
    Igor Afanasyev
    @iafan
    @timurgilfanov right. I realized I moved the Dockerfile into a subdirectory, but this messes with the COPY command. Will probably move that back. Your solution is correct. My command was docker build --no-cache -t serge -f docker/Dockerfile ., ran from the root Serge folder. (I kept no intermediate cache and explicitly tagged the image as serge)
    Igor Afanasyev
    @iafan
    @timurgilfanov for now, I decided not to move Dockerfile around, but instead add some more documentation. Maybe you will find some other parts of it useful for your case: https://github.com/evernote/serge/tree/master/docker
    gabrielgambarini
    @gabrielgambarini
    Hello everyone.
    I can't find in the documentation any parsing for excel files and different sheets inside spreadsheets . Any idea how this is done?
    Igor Afanasyev
    @iafan
    @gabrielgambarini Excel spreadsheets are rarely if ever a destination file format for translation. Do you happen to have them currently to just gather strings from engineers? How does your current localization process look like?
    gabrielgambarini
    @gabrielgambarini
    I work for an LSP company, receive excel sheets from publishing houses and from content producers. The best way would be to access it directly from their system for sure. But in reality, tables are 90% of the daily workflow. We receive all sorts of formats of tables, excel files, google spreadsheets, CSV ... in all sorts of formats with the target languages being columns, or sheets, or a segment of the same sheet but in different rows. So actually processing those tables is a big bottle neck in the process.
    Igor Afanasyev
    @iafan
    @gabrielgambarini the whole idea behind Serge is to simplify and automate the process of content delivery between the client and an LSP, exactly to remove those inefficient intermediate file formats. That’s why Serge doesn’t have a parser for Excel — in a proper workflow there’s simply no place for Excel. However, it is possible to parse Excel files, extract text from cells and put translations back in (I wrote some similar parsers before), but one needs to write such a parser. Feel free to PM me if you want to provide more details on your process, and whether you want to to try to ‘convert’ some if your clients to using an automated process.
    gabrielgambarini
    @gabrielgambarini
    @iafan Thanks! I will keep researching and I will contact you with some questions for sure. =)
    Igor Afanasyev
    @iafan
    @/all it's official: Serge 1.4 is out. Thanks everyone for your contributions!
    https://github.com/evernote/serge/releases/tag/1.4
    The new version has been uploaded to CPAN as well; will be indexed in a matter of next couple of days.