sillsdev / web-languageforge Goto Github PK

From @makerling :

Hi, I'm preparing for a workshop session on Languageforge and in my preparations I noticed a bug that occurs relating to sort order. I'm not sure if improvements have been made to the issue above, but I'm posting it here as it appears to be similarly related.

Steps to reproduce: (My username on languageforge is bluedragon). I created a new project from a FLEx lift file of approximately 1242 entries. The initial sort order is OK. I created a new entry and then clicked back to list. This time the order is jumbled and in no particular order. When I try to create another entry, the left hand side column list of entries shows the same jumbled order. I tried clicking around several areas as well as logging out and logging in to see if the Dictionary Download triggered the proper sort. I also tried Chrome, Chromium, Firefox to see if there were any changes. I cannot get the original sort order back. I also tried creating a new project, did not import a .lift file and manually created 4 entries. After creating those 4 entries with only ASCII characters, it still jumbles the sort order (see attached image).

Click for details
hasRight method cannot be called directly from ProjectModel line 157 /var/www/languageforge.org/htdocs/models/ProjectModel.php
sf.php line 713, RightsHelper->userCanAccessMethod(lex_configuration_update, Array)

JsonRpcServer.php line 67, f->checkPermissions(lex_configuration_update, Array)

api.php line 18, JsonRpcServer::handle(sf, CI_Output)

Happened when exporting Lift file. After I created folder to export to.

Ian

in the Test-Farsi project, search for fire (آتش) click on that entry in the search results.
In the list next to the entry there is a large gap in the entries. The entry above 'fire' is 'spittle' when actually in my FLEx project there are about 2 pages of entries between these two. See image attached.

I created a new project, and added one word, and set its Part of Speech to Verb (v). I then went to the Dictionary Configuration settings, on the Option Lists tab. I edited the Part of Speech list by adding a new category called Danger ahead (danger) -- the precise name doesn't matter as long as it's a new category. The newly-created category, at the bottom of the list, had a delete button next to it. BEFORE clicking Apply, I dragged that category up to the middle of the list. BUT the delete button stayed where it was, at the bottom of the list, and I was able to click on it to delete the Verb (v) category! The word I had entered, whose part of speech had been Verb (v), still showed as Verb (v) at first, and seemed to stay that way -- but at some point when I refreshed the page (without having ever edited any part of the word), its part of speech had switched to Unknown. (The Mongo database record, though, still contained "partOfSpeech" : { "value" : "v" }).

Here's a series of screenshots showing the steps to reproduce this bug. The mouse cursor isn't visible, but it should be obvious what I clicked on: the plus sign to add the new part of speech, then the drag bar to move the new entry, then the delete button to delete the wrong option.

What should have happened, of course, is that the delete button should have moved along with the newly-created option that it was "attached to".

This may be related to #27, as it shows similar symptoms.

I created a new, empty Dictionary project using French (fr) as the primary language. I then added an input system that's a variant of French using the Thai script (code fr-Thai, and I picked the abbreviation thfr for display in LanguageForge). I then clicked Apply, refreshed the configuration page so the new input systems would be picked up, and clicked on the Fields tab. I then created a new custom field at the Entry level, with the type Multi-input-system Text. The result was:

Note the checked, but blank, input system.

I suspect that creating a new variant input system (as I did) has nothing to do with the bug and that it would have happened even if I'd just had French and English in the project, but I haven't yet tested that scenario.

As Owner/Manager I was adding users. I added one user by just adding his name. A popup appeared with the associated password. After I copied down the password, I hit the add new user again while the previous name was still in the text box. It added the name again to the project but created a new password.

I should not be able to add the same name twice.

When I search for a word a pop-up comes up, but there is no way to scroll through all the results.

The search does not work with unicode accented vowels. I will check for compatibility with other characters. Glottal stops work fine.

It is easy enough to just search for the non-diacritic parts of words.

A further look reveals that if I copy and past directly from the entry, I get the accented a created as a combined diacritic and the search works.

Thanks!
Elliot

To reproduce:

• Create new project
• Click on a language (say, Thai)
• Click "More" and choose "Add a variant of Thai"
• Get error "Input system for Thai already exists"

What should happen:

New input system is created, with name "Thai (Variant 1)" or something, and the temporary name in parenthesis will be replaced with the actual variant name once the user has assigned one (and also probably specified a few variant options).

User reported: Under the "source" field we have the ampersand (&) and it is not showing as that but as "&".

In my Farsi project, the entry order seems to be messed up. This is the order I expected. (starting with the entries that have something in the fa writing system)

The third one should have an equals sign (ا=)

From LSA 2017 user Elliot:

Most desired features from talk are a list of active comments and more detailed activity log.

From Ron's email:

I tried to create a project named ‘test’ with code ‘tst’ and I got this error:

localhost:27017: file /var/lib/mongodb/sf_tst.ns open/create failed in createPrivateMap (look in log for more information) line 296 /var/www/languageforge.org/htdocs/models/mapper/MongoMapper.php

MapperModel.php line 126, MongoMapper->write(models\languageforge\lexicon\LexOptionListModel, )

The infinite scroll functionality seems to have regressed, no loading more entries when you scroll to the bottom of the list.

From LSA 2017 user Elliot

Api\Model\Languageforge\Lexicon\LexEntryModel(5849cfa289b1c7de24382825) doesn't exist line 48
/var/www/languageforge.org/htdocs/Api/Model/Shared/Mapper/MapperModel.php 
LexEntryCommands.php line 56, 
LexEntryModel->__construct(Api\Model\Languageforge\Lexicon\LexProjectModel, 5849cfa289b1c7de24382825)

Sf.php line 704, LexEntryCommands::updateEntry(58336c118b21477774361001, Array, 5717c521863f1cc81056541b)

I'm trying to import a project from landuagedepot, buy it doesn't work. The credential are correct, but is says: The username or password isn't valid on LanguageDepot.org. I've tryed with languagedepot and it works there.

Over in FLEx I entered an example sentence, but after I did S/R, in LF it looked like the Example was empty. In fact, on LF it was showing my two analysis languages for the Example, not the Vernacular language. Once I went to Fields and adjusted it, what I entered in FLEx appeared fine.

Then I tried with another brand new project (again with two ana languages). When I sent the project from FLEx to LD for the first time, there was only one entry, and no example entered. On LF, I got the project from LD. Then I clicked the orange "Add Example Sentence" in LF. Now it shows me the Example field, but only in the Ana languages, not the Vern language. It should default to showing the Vern language(s) for Example.

When I create a brand new LF project (not a S/R project), it comes up fine: the default for Example is the vern language.

Projects used for this test:
Indonesian Test
Test Chinese
Test French (LF only, not a S/R project)

in my Farsi project, for some reason when I look at the configuration for Citation Form. There is a blank writing system. See image below:

I added the wrong word list to my Borana project (i.e. a word list from a another language) by mistake. I would like to delete all the entries and start from scratch. How can I do that?

After the first initialization of test-pes-flex, I'm seeing entries with the POS showing as unknown. An example entry is the word glossed 'spittle' or the word glossed 'fire'. In the FLEx project on my machine these are showing as Noun.

PHP exceptions can be thrown from not escaping < and > in TsStrings of custom multi paragraphs

e.g. Someone sends a link to their dbe but you are not a member.

I cannot input any text into the Example field.

username: pektiong
project: Xnan

In FLEx I put my analysis language glosses in the gloss field, but I cannot get this field to display in the Language Forge. I realise that WeSay defaults to putting meaning data in the definition field so perhaps Language Forge uses definition for the meaning data, I don't know, but I'm not comfortable with putting glosses in the definition field. Perhaps there should be a way to tell Language Forge to look in the gloss field for meaning data.

See e.g. from GAB.

It seems that when multiple input systems are used for the word and meaning fields, only the first input system will show up in the list view. Is it possible to display data from all input systems of a field?

username: pektiong
project: Xnan

Some edits in custom line fields are not being preserved after Send/Receive.

-Editing custom single line all or custom single line ref, saving, then Send/Receive --> the field is refreshed with the original unchanged content
-Clearing a custom single line, then Send/Receive --> the original custom single line returns.

From the test plan scenario as below

Edit conflicts: FLEx edits entry, LF deletes entry

1. Record the FLEx project identifier that is used for this test scenario:
2. In one browser, login to Language Forge (LF) as the first user (Owner).
3. Browse to the LF project that corresponds to the FLEx project identifier.
   "4. If the Send/Receive status (next to the ""Send/Receive"" button) says ""Un-synced"", click the ""Send/Receive"" button.
   Wait until the status reports ""Synced""."
4. In FLEx, edit an entry, and do Send/Receive.
5. Verify the edits are retained.
6. In LF, find the same entry that FLEx edited above, and delete the entry. Click "Save Entry" (or wait for autosave).
7. Click "Send/Receive" from the Owner. Wait for S/R to complete.
8. Verify the edited entry remains (is not deleted).

**After step number 9, the edited entry get deleted instead of retain.

From the test plan "https://docs.google.com/spreadsheets/d/1aKKigd8FT-ak8VSfONzVgelxogcyVy9LB50ZpFgj6Eg/edit#gid=876420129" - row number 124.

From LSA 2017 user Elliot

There seems to be some lagging for some users. Anything you can think of to optimize their experience a bit?

from Susanna

in my Persian project, if I search for the English word 'book', of the 15 words that match, none of them is 'کتاب'. I expected this word to show up since it matches wholly the search term. I kind find the word without a problem if I search for 'کَتاب'.

I was surprised to see a gloss line for Swedish in my Farsi project. That wasn't part of my FLEx project. I had previously imported my Swedish project. Maybe it picked it up from there.

In my Swedish project only one of the custom fields at the sense level shows up (Target Equivalent), the other called Target Sense Number doesn't show up. In my Farsi project, however, I see two custom fields I created at the sense level. In both projects, I'm not getting the data that should be in those fields. In the Farsi project I see data for the entry level custom fields.

When uploading an audio file for SF Community Checking app, get the following error:
Undefined index: filename line 29
/var/www/host/ScriptureForge/sfwebchecks/src/controllers/upload.php

If I create a new Dictionary project with the primary language as Thai (th), it results in a blank language code for new words in the lexicon, and the dictionary configuration shows the Thai writing system as either blank or non-existent. (E.g., click on Dictionary Configuration, and the Input Systems tab shows only English and no other systems. Click over to the Fields tab and there are two writing systems, "" and "en". The blank one is the one that should be Thai).

This doesn't happen with any other primary languages (as far as I've determined), just with Thai (th).

See graphic below:

This on the opening page and x is on the right side where RTL text gets typed.

Steps to reproduce: This behavior appear when follow scenario 2 from test plan doc on “S/R Testing 0.9.42” tab
LanguageDepot_Project: India TestLangProj 05 ID: test-inda-tlp-05
Prerequisites:
LF 0.9.42 and FLEx “India TestLangProj 04” project have a picture in an entry.

FLEx User: Window 8.1 OS

1. Select an entry. the entry have picture and caption field have name
2. Delete that picture and S/R sync to LanguageDepot

LanguageForge User: Window 7 OS

1. Select an entry(Which have picture and used/edited by FLEx user)
2. Modified the “Picture Caption” name in the field.(changed to en: English)
3. S/R sync to languageDepot
4. After sync is completed. There is no picture appear and caption field have name which modified by LF user.

FLEx User: Window 8.1 OS

1. S/R sync to languageDepot. Here, There is no picture appear in an entry but picture caption appeared. So, I just clicked the Picture field on label to lead green crash.

Please refer screenshot for this in 31st August 2016 google doc. the issue number is 1 from below link.
https://docs.google.com/document/d/1uetHFquLX5R6WI91GJjYQVD4s4JrS0CVQwPJpl6qsM4/edit

In my Swedish project I have a word that is just one letter 'i'. When I search to try to find it, it is not in the results at all. Some kind of preference needs to be given for short words that begin with the search letter. In FLEx this is also a problem, but at least it shows up as the 4th result. Thanks, Ron L.

It doesn't appear that the syncing on the LF side is working at all for me. I've tried it several times. Here's the latest error.

Rewording this issue since @rmunn recently fixed the original one.

I noticed that if an entry has an example sentence and translation of the example sentence, only the example is displayed, not the translation of the example sentence in the "List View".

Also it seems other fields such as Grammar Note, Anthropology Note, Summary Definition among others are ones that I noticed the same behavior (not displaying in the List View). Is there an ability to configure the "List View" of the lexicon, or is this a bug? It would help the user to view the data much easier if the List View is configurable or at least displays everything entered.

Thanks,

Ken Keyes' projects all have WAV files in FLEx. I'm guessing WAV files are what most people have in their projects, and we should support playback and upload of WAV files. My current understanding is that we only support MP3.

From LSA 2017 user Elliot:

I had thought some custom fields were working, but really just they were some default fields that Fieldworks calls custom. In fact, I had a custom field made in Languageforge disappear after a sync. I plan to just hold extra metadata in the existing notes fields.

The Sena 3 FieldWorks sample project, which is no longer part of the default FieldWorks installer but is an optional download, won't import on LanguageForge. The message "Please wait. Importing Sena 3.7z..." just sits at the top of the page and never goes away, and the project is created but never populated with data.

Steps to reproduce:

1. Download Sena 3 sample project from above link
2. Load project into FieldWorks 8.1.4 by using Project Management -> Restore a Project...
3. Export a full lexicon from FieldWorks
4. Zip the resulting folder (I used .7z format, but .zip format should produce same results)
5. Create new project on local LanguageForge installation (or dev.languageforge.org, which also fails)
6. Select Sena 3.7z file for initial data import
7. Import process never completes

In this test of a conflict scenario:

1. FLEx user deletes a picture, does Send/Receive
2. LF user edits the caption to the same picture, does Send/Receive
3. The file deletion of the FLEx picture wins, and the LF caption edit wins.

This yields a broken picture link on LF (blank non-existent fiilename), with an edited caption.

When FLEx to FLEx does this scenario, no picture is displayed, but the caption remains

From LSA 2017 user Elliot

I’m getting a <br/> added to the end of some entries after they are edited in languageforge and synced to FLEx.

Sometimes after doing a Send/Receive in LF, an exception is thrown:
Undefined index: paragraphsHtml line 27

From @rmlockwood

I have a Swedish FLEx project. I had a project created for it on
languagedepot today. I did the initial S/R to it and then I created a new LF
project and imported the Swedish project from languagedepot. When I went to
open the project I got this:

No vernacular words showing. When I examined a record I saw that for some
reason the Thai writing system was active for my words and not Swedish. I
don't recall ever having Thai as one of my writing systems for this project.

I won't touch the LF project so you guys can investigate.
Thanks for looking into this strange behavior,

Trying to import zipped LIFT file. (11MB) It never resolves after 12 hours of waiting. Next time I open project "Server Status Code 500"